Journal of Japan Association for Earthquake Engineering, Vol.2, No.1, 2002 LAS COLINAS LANDSLIDE CAUSED BY THE JANUARY 13, 2001 OFF THE COAST OF EL SALVADOR EARTHQUAKE Kazuo KONAGAI1, Jörgen JOHANSSON1, Paola MAYORCA1, Tetsuro YAMAMOTO2, Masakatsu MIYAJIMA3, Ryosuke UZUOKA4, Nelson E. PULIDO5, Freddy C DURAN6, Kyoji SASSA7 and Hiroshi FUKUOKA7 1Dr. Eng., Prof. and PhD candidates, repectively, Ins., Industrial Science, Univ. of Tokyo, Tokyo 153-8505, Japan, [email protected]. 2 Dr. Eng., Prof,, Dept.,Civil Eng., Yamaguchi Univ., Yamaguchi 755-8611, Japan. 3 Dr. Eng., Prof,, Dept., Civil Eng., Kanazawa Univ., Kanazawa 920-8667, Japan. 4Dr. Eng., Assoc. Prof., Dept., Civil Engineering, Tohoku Univ., Sendai 606-8501, Japan. 5 Dr. Sci, Earthquake Disaster Mittigation Research Center , Hyogo 673-0433, Japan. 6Dr. Eng., Dept., Civil Engineering systems, Kyoto Univ., Kyoto 606-8501, Japan. 7Dr. Eng., Prof. and Assoc. Prof., respectively, Disaster Prevention Research Inst., Kyoto University, Kyoto 611-0011, Japan ABSTRACT: El Salvador was struck by two devastating earthquakes within a month. The first quake of Jan. 13, 2001, which was centered off El Salvador's southern coast, damaged and/or destroyed nearly 108,000 houses, and killed at least 944 people, including hundreds of residents buried in a huge amount of soil slipped down Las Colinas mountainside in the city of Neuva San Salvador (Santa Tecla). This report outlines the findings obtained through the reconnaissance by the JSCE team and laboratory tests that followed it. Key Words: El Salvador earthquake, landslide, pumice INTRODUCTION Nearly 800 volcanoes are active today or known to have been active in historical times. Of these, more than 75 percent are found in the Pacific Ring of Fire, the belt, partly coinciding with the young mountain ranges of western North and South America, and the volcanic island arcs fringing the north and western sides of the Pacific basin, includes Japan, Peru and El Salvador. The locations of the great majority of earthquakes also correspond to this belt. In these countries, landslides and debris flows are serious threats because of their extremely large travel distances. In the 1970 Peru Earthquake, for example, the huge soil mass of a couple of tens million m3 was initiated at the top of the Mt. Huascaran (6,700m EL), and ran about 4 km down to Yungay killing more than 20,000 people. El Salvador, one of the smallest and most crowded nations in Central America, extends about 240 1 kilometers westward from the Gulf of Fonseca to the border with Guatemala. This country was struck by two devastating earthquakes within a month. The first quake of Jan. 13, 2001, which was centered off El Salvador's southern coast, damaged and/or destroyed nearly 108,000 houses, and killed at least 944 people. One of the most spectacular aspects featuring heavily in the January 13 earthquake was the damage inflicted by landslides. Among them, Las Colinas landslide was the most tragic. A huge amount of soil mass (about 200,000 m3) was thrown off the rim of a mountain ridge rising south behind Las Colinas area of Nueva San Salvador (Santa Tecla), and flushed many houses and therefore more than 500 lives to death. This report outlines some important features of this landslide and discusses possible measures for reducing preventive loss of human lives and social disruptions. OVERVIEW OF LAS COLINAS LANDSLIDE A mountain ridge called El Balsamo rises south behind the city, and the slope failure took place in its northern side (Fig. 1). Other small landslides and cracks are also found along the mountain ridge. Just behind the top scar is a flat area slightly slanting southwards, and there were two unfinished brick houses, which were supposed to become a school. A coffee plantation spreads along the ridge keeping the soil moistened in places. Zone 1 Zone 3 Zone 2 Fig. 1. Bird’s eyes view of the Las Colinas landslide Fig. 2. Surveyed Slope plotted with surfer with green dashed perimeter. 2 The Las Colinas sliding surface was surveyed with a laser based theodolite (Laser Ace 300), connected to a portable computer. The theodolite has a built-in digital compass, and with its laser beam, it calculates the azimuth, dip angle, and horizontal distance to a point. The exposed failure surface can be roughly divided into three zones (Fig. 2). The uppermost zone (Zone 1) is a hollow of about 100m in diameter, which was caved in some 20-30 m from the original ground surface. South beyond the hollow, there appears a steepest slope (Zone 2), which becomes gradually gentle as it comes close to the toe. In Zone 3, the slid soil mass and flushed houses had been removed already, and the cleared area was totally white with disinfectant (see Fig. 1). DETAILED FEATURES OF THE SLOPE Uppermost main scarp (Zone 1) Two photos (Fig. 3) of the same soil layers appearing on the west half of the uppermost scar were taken on January 14 and February 4, respectively. Dark and blight colored stripes appearing on the scarp shows a stratified soil profile with the top lapilli tuff layer of about 2.0 m thickness overlying a pair of two differently colored (ocher and white) pumice layers of about 11.9m thickness. Some bedding planes separate the lapilli tuff into some sub-layers of varying thickness (10cm to 25cm). Dark brown color of soil indicates that the soil is moistened, and this pair of photos shows that the dark colored stripes were steadily thinning, in other words, exposed soils were drying. This fact may evidence that the intact soils along the slip surface were wet before the event. Beneath the dark and wet soils, there appear white and/or ocher colored pumice soils. They were also moistened and weakly cemented. The pumice is easily broken into small pieces or into powder just by rubbing together with fingers. Broken fragments of this pumice have sub-angular or angular shapes. Some pieces were medium-sized (1 mm × 4 mm × 7 mm for example) and some were fairly large reaching 12 mm × 25 mm × 35 mm (Fig. 4). As shown in Fig. 5, cracks and joints were observed even on an intact part of pumice. (a) January 14 (b) February 4 Photo. by Mr. Jose Antonio Rivas Fig. 3. West slope of main scarp 3 Fig. 4 A piece of lapilli tuff Fig. 5. Cracks and joints on the exposed intact pumice Table 1 Crack openings along Line A-A’ in Fig. 6 (see next page) No. Distance from A Vertical offset Opening Cumulative Depth (cm) (cm) (cm) opening (cm) (cm) 1 180-240 10 60 60 122 2 390-410 7 20 80 65 3 460-471 5 11 91 60 4 580-581 0 1 92 20 5 590-593 0 3 95 20 6 780-782 0 2 97 20 7 833-834 0 1 98 20 8 1215-1217 0 2 100 30 9 500-1502 0 2 102 20 10 1525-1535 0 10 112 20 11 1841-1859 0 8 120 50 12 2234-2249 0 5 125 20 Total 43 points were marked along the perimeter of the scar using a GPS receiver (Fig. 6, next page). Behind the scar, there were extensive cracks running in almost west-east (See photos in Fig. 6). Crack openings (12 visible cracks; Table 3.2.1 and Figure 6) were measured along Line A-A’. Total 1.25 m’s opening was reached over the 22 m’s distance of Line A-A’. Namely, about 5% average strain was induced within the soil behind the scar. Extensive cracks were found along the ridge crest even in areas that did not slide (see Fig. 14); the cracks can certainly cause further slides. 4 Cracks appearing on the top terrace behind the scar (Zone 1): Cracks map is shown below. A' -20 A 0 Microtremor N 20 Reference point (Laser rangefinder) Distance in south-north (m) in south-north Distance 40 40 20 0 -20 -40 -60 -80 Distance in west-east (m) 0 A A' -2 01020 Distance (m) Fig. 6. Perimeter of scar, Las Colinas Landslide Fig. 7. Crack at the top terrace (photographed on Jan. 14 by Mr. Jose Antonio Rivas) 5 A pair of photos (Fig. 7 by Jose Antonio Riva, Geotérmica Salvadoreña) was taken on January 14, a day after the earthquake. Both pictures show that edges of cracks were lightly dusted with fine sand. This fact may evidence that the landslide caused some soils to liquefy though the underground water level did not seem to be substantially high from the field observation. There was no written material available showing the possible underground water level. There are some downward streaks remaining on the bottom of the uppper hollow (arrow in Fig. 8), and broken pieces of a mortar-block fence were caught on the soil mass remaining on the outer edge of the hollow (arrow in Fig. 9). These may evidence that most soil mass, which originally fitted into the hollow, seems to have been pushed forward, and flown down the lower steep slope. Along the streaks, a number of broken pieces of lapilli tuff and some uprooted trees were found. The largest piece of the lapilli tuff remaining there was 88 cm×64 cm×14 cm. A piece of mortar block fence Fig. 8. Lower part of the hollow: A piece of mortar block fence (within the circle) is caught on the small amount of the soil mass stopped at the lower edge of the hollow. Fig. 9. Fallen block fence 6 Steep slope in the middle (Zone 2) Fig.